Revert "[BugFix] Support setting tp=1 for the Eagle draft model to take effect (#5519)"(#5902)

This reverts commit d886b81971. it breaks pd function

- vLLM version: v0.13.0
- vLLM main:
bde38c11df

Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com>

tests/ut/spec_decode/test_eagle_proposer.py

@@ -27,8 +27,6 @@ class TestEagleProposerInitialization(TestBase):
         self.vllm_config.model_config.dtype = torch.float16
         self.vllm_config.model_config.max_model_len = 2048
         self.vllm_config.model_config.uses_mrope = False
-        self.vllm_config.parallel_config.tensor_parallel_size = 1
-        self.vllm_config.speculative_config.draft_tensor_parallel_size = 1
         self.vllm_config.speculative_config.num_speculative_tokens = 2
         self.vllm_config.speculative_config.speculative_token_tree = str([
             (i + 1) * (0, ) for i in range(2)
@@ -116,8 +114,6 @@ class TestEagleProposerLoadModel(TestBase):
         self.vllm_config.model_config.dtype = torch.float16
         self.vllm_config.model_config.max_model_len = 2048
         self.vllm_config.model_config.uses_mrope = False
-        self.vllm_config.parallel_config.tensor_parallel_size = 1
-        self.vllm_config.speculative_config.draft_tensor_parallel_size = 1
         self.vllm_config.speculative_config.num_speculative_tokens = 2
         self.vllm_config.speculative_config.speculative_token_tree = str([
             (i + 1) * (0, ) for i in range(2)
@@ -250,8 +246,6 @@ class TestEagleProposerDummyRun(TestBase):
         self.vllm_config.model_config.dtype = torch.float16
         self.vllm_config.model_config.max_model_len = 2048
         self.vllm_config.model_config.uses_mrope = False
-        self.vllm_config.parallel_config.tensor_parallel_size = 1
-        self.vllm_config.speculative_config.draft_tensor_parallel_size = 1
         self.vllm_config.speculative_config.speculative_token_tree = str([
             (i + 1) * (0, ) for i in range(4)
         ])
@@ -366,8 +360,6 @@ class TestEagleProposerHelperMethods(TestBase):
         self.vllm_config.model_config.dtype = torch.float16
         self.vllm_config.model_config.max_model_len = 2048
         self.vllm_config.model_config.uses_mrope = False
-        self.vllm_config.parallel_config.tensor_parallel_size = 1
-        self.vllm_config.speculative_config.draft_tensor_parallel_size = 1
         self.vllm_config.speculative_config.num_speculative_tokens = 2
         self.vllm_config.speculative_config.speculative_token_tree = str([
             (i + 1) * (0, ) for i in range(2)

tests/ut/spec_decode/test_mtp_proposer.py

@@ -42,9 +42,6 @@ class TestMtpProposer:
         config.model_config.max_model_len = 2048
         config.model_config.uses_mrope = False
         config.model_config.hf_text_config = None
-        config.model_config.hf_config = None
-        config.parallel_config.tensor_parallel_size = 1
-        config.speculative_config.draft_tensor_parallel_size = 1
 
         config.load_config = None